A simple sequencing index separated clonal from independent GI tumours far better than morphology

Design

Retrospective diagnostic accuracy study with validation cohort

Population

120 paired GI, HPB and peritoneal carcinomas from 60 patients

Primary outcome

Accuracy of clonal relatedness index vs phenotype for clonality

Effect

AUC 0.986; sensitivity 96.2%, specificity 100% vs 71.2% and 62.5% for phenotype

Investigators compared tumour-specific alterations in 120 paired gastrointestinal, hepatopancreaticobiliary and peritoneal carcinomas from 60 patients whose clonal status had been established clinicopathologically (Journal of Molecular Diagnostics, 25 September 2026). The clonal relatedness index is the proportion of shared alterations over the average total number of alterations in the two tumours.

Across 1,508 alterations, the median index was 0.85 in clonal pairs and 0.24 in non-clonal pairs (area under the curve 0.986). At a cut-off of 0.53, the index had 96.2% sensitivity and 100% specificity, compared with 71.2% and 62.5% for morphology and immunohistochemistry. In hypermutated tumours, morphology fell to 42.9% sensitivity while the index held at 85.7%.

Where a patient has two GI carcinomas and the question is metastasis versus a second primary, and sequencing has already been done on both, the index is a straightforward way to use those data.

• Consider a clonal relatedness index when two GI carcinomas have both been sequenced and morphology is equivocal.
• Be especially cautious with morphology alone in hypermutated (for example mismatch repair deficient) tumours.
• State in the report which method established relatedness.
• Note the small number of non-clonal pairs (8) when quoting specificity.

Why it matters

Calling a second primary a metastasis, or the reverse, changes stage and treatment.

Don't overread it

Only eight non-clonal pairs were studied, so the 100% specificity is imprecise.

The statistics, in plain English

An area under the curve of 0.986 means near-perfect separation in this sample. With so few non-clonal cases, a single misclassified pair would change specificity by about 12 points.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.